Bridgehead carboxy-substituted 2,3-diazabicyclo[2.2.2]oct-2-enes: synthesis, fluorescent properties, and host-guest complexation

Two novel derivatives of 2,3-diazabicyclo[2.2.2]oct-2-ene were synthesized, carrying a carboxyl ( 4 ) and a methylcarboxyl ( 5 ) substituent at the bridgehead position. The photodecomposition quantum yields (51% for 4 and 2.9% for 5 ) and fluorescence lifetimes (29 ns for 4 and 345 ns for 5 ) in water were determined. The higher photoreactivity and fluorescence quenching for 4 was attributed to its higher propensity to undergo photochemical elimination of nitrogen as a consequence of the presence of the radical-stabilizing carboxyl group in the α -position. The absolute photodecomposition rate constants of 4 became faster upon protonation (e.g., at pH 2), which contrasted anticipated substituent effects on the C–N bonds strengths. The complexation behavior of both acids, 4 and 5 , with tetrakis( N , N '-dimethylammonio- methyl)tetrahydroxycalix[4]arene ( 17 ) and β -cyclodextrin was studied by fluorescence and induced circular dichroism, respectively, in order to evaluate their suitability as potential sensor systems for anions and analytes with hydrogen-bonding abilities, respectively. The binding constants of 4 and 5 with the calixarene 17 were unexpectedly small (< 2000 M –1 ) which was attributed to reduced Coulomb attractions as a consequence of the 1,3-alternate conformation which this host adopts in water. The binding constants towards β -cyclodextrin were also low at pH 7.0 (< 120 M –1 ), which was attributed to the low hydrophobicity of the anionic form of the guests; in line with this interpretation, the binding constants with β -cyclodextrin increased at pH 2.0, by about one order of magnitude.


Introduction
2][3] They have remained key intermediates in the synthesis of strained compounds 1,4 and in the mechanistic investigation of the behavior of biradicals, [5][6][7][8][9][10][11] with some seemingly never-ending mechanistic disputes like the double-inversion peculiarity of 2,3-diazabicyclo[2.2.1]hept-2-enes. 12,137][28][29] Herein, we describe the DBO carboxylic acid 4 and the DBO acetic acid 5.We synthesized these derivatives for two main reasons: First, we wanted to broaden the repertoire of available derivatives for subsequent biolabeling, e.g., N-hydroxysuccinimid (NHS) esters are popular compounds for post-column peptide labeling of amino groups, and secondly, we were interested in the effects of a negative charge on the complexation of the DBO chromophore with macrocyclic host molecules, with the ultimate perspective to design refined sensor applications. 36Herein, we describe the synthesis of 4 and 5, the photophysical and photochemical characterization of both acids, including some unexpected peculiarities, and their complexation behavior with β-cyclodextrin and positively charged calix [4]arenes.

Synthesis
The synthesis of DBO was first reported in 1962 by Cohen and Zand 37 and since then many derivatives have been prepared owing to their interesting photophysical properties.In fact, Adam 38 and others [39][40][41][42] already studied more complex polycyclic derivatives carrying bridgehead carboxyl substituents, e.g.derivative 6, which can be obtained under mild acid-catalyzed conditions from the reaction of 4,5-dihydropyridazines with electron-rich alkenes (the Sauer-Hünig route to polycyclic azoalkanes). 43,44Moreover, Engel and coworkers synthesized the DBO derivative 7 carrying a carboxyl group in the 5 endo position, 41 which is obtained under the harsh hydrolysis conditions of the urazole route (heating in concentrated base). 41In contrast, the simple bridgehead carboxyl derivative 4 has remained unknown.Unfortunately, the known derivatives, like 6 and 7, are entirely unsuitable for the desired applications in biomolecular and supramolecular chemistry.On one hand, DBO derivatives are frequently preferred due to their high hydrophilicity and small size, which is not fulfilled for derivative 6; in addition, the presence of two functional groups in 6 (rather than a single carboxyl group) is most undesirable for biomolecular labeling purposes.On the other hand, derivative 7 is obtained as a racemic mixture, which would lead to the formation of two diastereomers in the labeling of biomolecules, in particular polypeptides, or in the molecular recognition with chiral hosts like cyclodextrins. 18,23Such diastereomeric complications need to be avoided at all cost, in particular in mechanistic investigations.To by-pass the above problems, while maintaining high water-solubility, we have previously synthesized and comprehensively employed the derivatives 2 and 3. 14,[23][24][25][26]30 Several attempts to oxidize the hydroxymethyl DBO 2 directly to the bridgehead carboxy-substituted derivative 4 by several mild oxidation methods failed, such that we needed to resort to the conventional synthesis route based on cycloaddition with 4-N-methyl-1,2,4-triazolin-3,5-dione (MTAD, Scheme 1). 1  The cycloaddition with MTAD required cyclohexa-1,3-dienecarboxylic acid (9), which had been previously obtained by a different route. 45We obtained 9 in excellent yield from the readily accessible 44 cyclohexa-1,3-dienecarbaldehyde (8) by ambient-temperature oxidation 46 with freshly precipitated silver hydroxide in a THF/water-mixture; as a minor oxidation by-product, benzoic acid (< 5%, depending on reaction time and concentration) was identified.Since the latter was not expected to interfere with the following cycloaddition reaction with MTAD, we used the crude oxidation product.The cycloaddition with MTAD proceeded smoothly at ambient temperature, with the usual white-to-pink color change as a convenient indicator for complete conversion; however, the urazole 10 precipitated partially due to its limited solubility in the dichloromethane solvent.Benzoic acid, which presented an impurity from the preceding oxidation step, was removed by the subsequent recrystallization of 10 from ethanol.The low solubility of 10 in ethanol precluded also the subsequent use of this solvent in the hydrogenation step.Dioxane was selected as an alternative, and the hydrogenation (ambient temperature, 1 atm) proceeded quantitatively as well.The product was extensively purified by several recrystallizations from i-propanol, because the purity was deemed an essential factor for the success of the subsequent hydrolysis to produce 4.

O
Commonly, oxidation of the hydrazine, as the initial urazole hydrolysis product, to the azoalkane can be effected in-situ by carrying out the hydrolysis under air. 47To prepare azoalkane 4, however, we found a cleaner reaction when we separated the hydrolysis step from the oxidation by carrying it out under inert atmosphere.In addition, the hydrolysis time needed to be prolonged significantly (2 days) to affect higher conversion in larger scales.Subsequently, the hydrolyzed product (hydrazine) was given 3 days to oxidize under air.The highly water-soluble product 4 was obtained by acidification and recrystallization from ethyl acetate in 42 % yield.We suspected first that a thermal instability of 4 was responsible for the moderate yield in this reaction.Note that azoalkanes with tertiary α-carbons and at least one radical-stabilizing substituent (with azo isobutyronitrile, AIBN, being the prototypal example), are known to decompose readily even at only slightly elevated temperature. 48,49However, control experiments showed that compound 4 was thermally stable for at least 72 h even in water at 95 °C (no decomposition observed by UV).
The DBO acetic acid 5 was prepared according to Scheme 2. Nucleophilic substitution of the previously reported 30 tosylate 12 is known to be unfavorable due to the steric hindrance imposed by the neopentyl-type position; in the preparation of 3 (azide substitution with subsequent reduction) the use of sodium azide in HMPT at elevated temperature for 14 h was required. 30For the preparation of azoalkane 5 (cyanide substitution with subsequent alkaline hydrolysis), the higher nucleophilicity of cyanide 50 allows one to avoid the highly toxic HMPT, because already the use of an excess of sodium cyanide in dry DMSO at 110 °C for 24 h led to the formation of nitrile 13.The subsequent hydrolysis to 5 proceeded smoothly.
In sharp contrast stands the carboxylic acid 4, which has a relatively short fluorescence lifetime (τ = 29 ns at pH > 5, Table 1).Such short fluorescence lifetimes are already known for some DBO derivatives, namely those carrying carboxylic acid ester, phenyl, or vinyl groups in the bridgehead positions. 38,53A specific example is the 1-phenyl derivative 14.It was concluded that these substituents decrease the activation energy and accelerate the rate of nitrogen extrusion by stabilizing the incipient carbon radical site. 38,48,53,56This affects also the quantum yield of photodecomposition (φ r ), which we have consequently determined 57 in water as solvent with reference to the parent 1 (φ r = 0.003). 17We found that the photodecomposition quantum yield was only slightly enhanced for 5 (φ r = 0.029), but much higher for 4 (φ r ca.0.5).The results are in accord with literature data, which predict 10,56,58,59 a less efficient radical stabilization by a carboxylic acid substituent than, for example, by a phenyl substituent.In fact, azoalkane 14 decomposes with unit quantum efficiency (φ r = 1.0, in benzene). 53,56,58 The resonance-stabilizing effect of a substituent is also reflected in the product distribution of azoalkanes photolysis.The commonly observed photoproducts of DBO derivatives are 1,5hexadienes and bicyclo[2.2.0]hexanes; the yield of the former photoproduct is assumed to increase with the delocalizing ability of the substituents, because of the general driving force to form a larger conjugated system. 53 1H NMR analysis (without isolation of the individual products) of the photolysis products from 4 revealed 2-methylenehex-5-enoic acid 15 60,61 and the bicyclo[2.2.0]hexane-1-carboxylic acid 16 62 in a product ratio of ~90:10.For comparison, the conjugation with a phenyl group in 14 produces solely 2-phenyl-1,5-hexadiene (in benzene), 53 while the parent 1 produces ca.70 % hexadiene and 30% bicyclo[2.2.0]hexane (in water). 17

16
The photophysical properties of azoalkanes 4 and 5 were determined by UV absorption as well as steady-state and time-resolved fluorescence spectroscopy in aerated water at pH 7.0 (Table 1).The fluorescence lifetime of 5 (345 ns) can be considered as "unquenched" in comparison to the parent 1 (325 ns). 17In contrast, the fluorescence lifetime of 4 (29 ns at pH > 5) is much shorter, which is an expected manifestation of its ca.15 times larger photodecomposition quantum yield (vide supra); for comparison, the 1-phenyl derivative 14, which eliminates nitrogen photochemically with unit quantum efficiency, is virtually nonfluorescent and has a very short (2.5 ns) fluorescence lifetime. 53The relative fluorescence intensities showed the same trends as the fluorescence lifetimes, as expected, while the UV absorption spectra revealed only very minor shifts as a consequence of the bridgehead substitution (Table 1).
The high water solubility of azoalkane 4 in both its protonated (carboxylic acid) and unprotonated (carboxylate) forms allowed us also to perform pH-dependent studies.In detail, we measured the quantum yields for photodecomposition (Φ r , Table 1) and the fluorescence lifetimes (τ, Table 1) at pH 7 and 2 in order to determine the absolute photochemical reaction rates (k r = Φ r /τ).We found that photochemical nitrogen extrusion in 4 proceeds 5-6 times faster at pH 2 (k r = 9.8 × 10 7 s -1 ) than at pH 7 (k r = 1.8 × 10 7 s -1 ).This reactivity pattern contrasts expectations from substituent effects on the bond dissociation energies of the C-N bond to be broken during the photochemical reaction.For example, C-H bond dissociation energies are higher for acetic acid (95.3 kcal/mol) than in the acetate anion (93.7 kcal/mol). 63This trend was also reproduced by semi-empirical calculations (UHF-AM1) for the two prototropic forms of azoalkanes 4, which revealed also a 0.9 kcal/mol lower C-N bond dissociation energy for the anionic form, suggesting again that protonation of the carboxylic acid group at pH 2 should slow denitrogenation, and not accelerate it as experimentally observed.Obviously, the photoinduced bond cleavage in 4 is governed by additional factors, presumably polar effects on the transitionstate.
The difference in photochemical reactivity accounts also directly for the strong fluorescence quenching observed for both steady-state intensities and lifetimes of azoalkane 4 when we lowered the pH from 7 to 2 (see Figure 1).The quenching was somewhat more pronounced in the steady-state (factor ca.10) than in the time-resolved (factor ca. 6) measurements.The fact that the largest change occurred near the typical pK a values of carboxylic acids (ca.4) demonstrated nicely that the protonation of the carboxylic acid group was responsible for the fluorescence quenching.For comparison, absorption spectra remained virtually unchanged down to pH 1.5, which confirmed that no protonation of the azo chromophore had occurred in the relevant low pH range; 15   The fluorescence quenching is clearly due to an intramolecular process, i.e., photochemical reaction.Intermolecular quenching was readily excluded by the fluorescence lifetimes, which were independent of fluorophore concentration in the range of 0.1-1 mM; in addition, we remeasured the fluorescence lifetimes of 1 in degassed acetic acid and observed rather long fluorescence lifetimes and consequently inefficient intermolecular quenching by carboxylic acid groups (τ = 220 ns, k q = 3.6 × 10 6 s -1 , this work, calculated with reference to an unquenched (gas phase) fluorescence lifetimes of τ 0 = 1030 ns), 64 as observed previously. 52However a deuterium ARKAT USA, Inc.
isotope effect of ~3.6 for the solvent-induced quenching by acetic acid (acetic acid-d1, τ = 490 ns, k q = 1.1 × 10 6 s -1 , this work) was also established, which suggests that a carboxylic acid substituent does nevertheless act as a weak quencher (hydrogen atom donor).Its main influence on the photophysical properties of 4 is, however, manifested in the acceleration of its photodecomposition, which reveals also an interesting pH dependence (vide supra).
Another interesting peculiarity of 4 is the observed difference in pK a values derived from the steady-state and time-resolved fluorescence titrations (see Figure 1).This variation is not yet understood, but presumably due to additional protonation equilibria in the excited-state.The pK a value as determined by time-resolved-fluorescence is ~3.1 in D 2 O as well as in H 2 O.However, the pK a value as determined by steady-state fluorescence is higher, by 0.6 units in H 2 O and by 0.9 units in D 2 O.With the photophysical data and the effects of protonation in hand, we proceeded to study host-guest complexation phenomena to evaluate the potential of 4 and 5 to serve as novel fluorescent probes.

Complexation with tetrakis(N,N'-dimethylammoniomethyl)tetrahydroxycalix[4]arene (17)
We have recently introduced a new supramolecular sensor system for organic cations (e.g., cholines and carnitines), which is based on 1 and the cation receptor p-sulfonatocalix [4]arene (18), which displays a binding constant (K a ) of 1200 M -1 at pD 7.4. 15As an extension of this work, the sensor system has been refined by using the amino derivative 3 (in its ammonium form) which allows for additional Coulomb attractions with the four sulfonato groups in the cone conformation of 18.This is reflected in a strongly increased binding constant (K a = 6 × 10 4 M -1 for 3 at pH 7.0), 65 which improves the sensitivity of the sensor system.Conversely, we were interested in the ability of the cationic calixarene 17 to complex the novel DBO acids 4 and 5 in their anionic form and thereby devise potential sensor systems for anionic analytes.The host-guest complexation experiments were performed at pH 5.5 where all ammonium groups of host 17 are expected to be protonated, but where the carboxylic acid groups of the azoalkanes (as guests) are deprotonated (Figure 1).Under these conditions, we found apparent binding constants by fluorescence titrations of 800 M -1 for 4 and 1900 M -1 for 5 (see Figure 2).Because the fluorescence quenching is partially related to a dynamic fluorescence quenching, 15,17,18,22 which is expected to be more pronounced for 5 due to its longer fluorescence lifetime (vide supra), the binding constants derived from these titrations need to be considered as limiting maximum values (< 2000 M -1 ).Unfortunately, independent UV titrations could not be performed in order to obtain more accurate binding constants, because the strong and broad UV absorption band of 17 in the range of 320-380 nm prevented a monitoring of the weak UV absorption of the azo chromophore. 3,15,17mportant to note, the estimated binding constants of 4 and 5 with host 17 (< 2000 M -1 at pH 5.5) are unexpectedly low compared to the binding of 3 with host 18 (K a = 6 × 10 4 M -1 ).Indeed, 1 H NMR experiments revealed that 17 does not adopt a flexible cone conformation as 18 does, 66 but instead a 1,3-alternate conformation in D 2 O (pD = 6.0).This contrasts the cone conformation reported for 17 in CDCl 3 67 and independently predicted by molecular modeling studies in water, 68 but is consistent with the 1,3-alternate conformation determined for the closely related tetrakis(N,N',N''-trimethylammoniomethyl)tetramethoxy-calix [4]arene in water. 69Consequently, we attribute the low binding constant of 4 and 5 with 17 to the 1,3-alternate conformation of this host in water, because Coulomb attractions are reduced in the 1,3-alternate conformation compared to those in a cone conformation, where all four charged aryl groups can interact with the guest.

Complexation with β-cyclodextrin
8][19][20][21][22][23] We have now expanded the studies to the DBO acids 4 and 5. Complex formation was followed by ICD (see Table 2), where spectral effects accompanying host-guest complexation are exaggerated compared to those observed by UV spectrophotometry, 18,21 and more straightforward to interpret ARKAT USA, Inc.
Interestingly, the binding constants of both 4 and 5 at pH 7.0 are very low (< 120 M -1 ), which contrasts with previously reported binding constants of bicyclic bridgehead carboxysubstituted alkanes (e.g., for the bicyclo[2.2.2]octane analogue of 4 a binding constant of 9.7 × 10 4 M -1 has been reported and 9.5 × 10 5 M -1 for adamantanecarboxylic acid). 70This can be rationalized by considering the high intrinsic water solubility of bicyclic azoalkanes compared to bicyclic alkanes, which lowers the hydrophobic contribution to the driving force of complexation with cyclodextrins. 22The importance of guest hydrophobicity was also confirmed by the fact that the binding constants increased by ca.one order of magnitude when the carboxylic acid group was protonated (e.g., at pH 2.0, Table 2).As previously observed for the two prototropic forms of the aminomethyl derivative 3, the noncharged form has a higher affinity to cyclodextrins. 22A difference by one order of magnitude appears to be characteristic, which manifests itself also in a host-induced pK a shift, i.e., the pK a value of carboxylic acids included by β-cyclodextrin is expected to be one unit higher than in their uncomplexed state. 16he signs and magnitudes of the ICD signals (Figure 3) allow conclusions about the geometry of the host-guest complexes to be drawn, namely on the relative orientation of the guest to the host ("co-conformation").The method has been exemplified in detail for bicyclic azoalkanes in previous studies, [20][21][22] and the assignments were based on the rules of Harata 71 and Kodaka. 72For example, for cyclodextrin inclusion complexes of bridgehead-substituted azoalkanes, we have assumed frontal inclusion geometries (with the bridgehead hydrogen protruding into the cavity and the substituent interacting with the upper rim).While a negative ICD signal is a priori expected for a frontal co-conformation, positive signals can be interpreted in terms of a significant deviation (tilting) from the frontal geometry, presumably due to specific interactions of the functional groups of the guest with the hydroxyl network at the upper cyclodextrin rim.Accordingly, the observed (weak) ellipticities (Table 2) may be tentatively interpreted in terms of an accurately frontal geometry for azoalkane 4 at pH 2 (negative sign), while for the other cases (positive signs) a tilted geometry is indicated (Scheme 3).The inversion of the ICD sign of azoalkane 4 at neutral and acidic pH (Figure 3) is reminiscent of the inversion in ICD sign previously observed for azoalkane 3 at neutral and alkaline pH. 22Variations in hydrogen bonding interactions at the upper cyclodextrin rim have been held responsible for these subtle changes in the complex geometry.pulsed diode laser (λ exc = 373 nm, λ em = 450 nm, ca.50 ps pulsewidth) for excitation.
Circular dichroism spectra were recorded on a Jasco J-810 (0.5 nm resolution, 25 accumulations) by using a cyclodextrin solution for background correction.Photoreactions (also for photodecomposition quantum yield determinations) were carried out in a Luzchem LZC-4V photoreactor equipped with 14 Hitachi FL8BL-B UVA lamps (λ max,em = 350 nm).NMR-spectra were recorded on a JEOL JNM-ECX400 working at 400 MHz for 1 H and 100 MHz for 13 C measurements.pH measurements were performed with a WTW 330i pH meter equipped with a WTW SenTix Mic glass electrode.Conversion to pD values was made on the basis of ref. 73 (+ 0.40 units).X-ray analysis has been performed on a Bruker X8 APEX II CCD diffractometer with kappa geometry using MoKα (λ = 0.71073 Å) radiation at 173(2) K. Semi-empirical computations were carried out by applying the AM1/UHF force field embedded within Hyperchem ver.7.1 (Hypercube Inc., Gainesville, FL).Synthesis 1-Hydroxycarbonyl-1,3-cyclohexadiene (9).The aldehyde 8 was prepared according to ref. 44 .25.5 g (0.15 mol) of silver nitrate were dissolved in 700 ml water and added to a stirred solution of 12.0 g (0.30 mol) sodium hydroxide in 700 ml water.To the resulting grey suspension 5.4 g (0.05 mol) of aldehyde 8, dissolved in 15 ml of THF, were added over the course of one hour.The mixture was stirred overnight at ambient temperature.After removal of THF under vacuum, the suspension was filtered and washed with hot water.After acidification with 2 N HCl the organic acid was extracted with diethylether.The combined ether extracts were washed with water and dried over magnesium sulfate.After filtration and removal of the solvent, a pale, yellow oil was obtained that crystallized in the refrigerator.Recrystallization from ethanol/water (50/50) yielded 6.2 g (100% conversion) of 9 as white to pale yellow crystals with identical physical properties to ref. 45 .The NMR spectra of the crude product showed various amounts of aromatized product (benzoic acid).1-Hydroxycarbonyl-4-methyl-2,4,6-triaza-tricyclo[5.2.2.0 2.6.]undec-8-ene-3,5-dione (10).A solution of 5 g (44.2 mmol) of MTAD in 100 ml dichloromethane was added dropwise to a stirred solution of 5.5 g (44.3 mmol) of the diene acid (9) in 100 ml dichloromethane at ambient temperature.After about half of the MTAD solution had been added, the reaction mixture started to become cloudy and a white precipitate was formed.At the end of the addition, the pink color remained permanently (unreacted MTAD).The suspension was vacuum filtrated and washed with ethanol at ambient temperature.The remaining white powder was recrystallized from ethanol.Evaporation of the combined supernatant and washing liquids yielded a further fraction of crude cycloaddition product which was purified in the same way.The combined fractions of white crystalline powder gave 10.3g (98%) 10. m.p.: 193 °C ARKAT USA, Inc. ]undecan-3,5-dione (11).10 g (42.15 mmol) of the olefinic acid 10 were dissolved in 200 ml dioxane in a normal pressure hydrogenation apparatus and 100 mg of 5% palladium on carbon catalyst were added.The hydrogenation was conducted overnight until no more hydrogen was consumed.The reaction mixture was filtered through filter paper and additionally through a 1 cm layer of silica to remove traces of fine carbon particles.After removal of the solvent under vacuum a honey-like residue was obtained, which became a glassy solid upon standing in the refrigerator.16.7 g (0.3 mol) KOH pellets were dissolved in 100 ml i-propanol and 5 g (20.09 mmol) of 11 were added under inert atmosphere.The mixture became clear upon heating to reflux.After about half an hour, a white precipitate formed (presumably potassium carbonate).Refluxing under inert atmosphere was continued for 2 days.The mixture was allowed to cool to ambient temperature and transferred into a beaker.The beaker was left open under air for 3 days (stirring during the first 24 h) to oxidize the intermediary hydrazine to the corresponding azo compound and to evaporate i-propanol.To the resulting white solid, 25% HCl was added dropwise (CO 2 -evolution!) under cooling with ice until a clear solution was formed with neutral pH.Further addition of HCl (up to pH ca. 1) afforded a white microcrystalline precipitate.The liquid was removed carefully with a pipette and the residue was washed twice with a small amount of distilled water to remove remaining KCl traces.After the solid was dried under vacuum a recrystallization from ethyl acetate with hot filtration was performed.Suction filtration and drying under air yielded 1.23g (42%) of 4 as white crystals.An additional slow crystallization from ethanol produced single crystals suitable for X-ray structure analysis.The resulting structure is shown in Figure 4. m.p.: 142 °C (dec.) 1 H-NMR (CDCl 3 ): δ 1.30-1.50(4H, m), 1.60-1.80(2H, m), 1.95-2.15(2H, m), 5.33 (1H, m), 6.20-7.80(1H, br) 13 C-NMR (CDCl 3 ): δ 20.9 (t), 24.5 (t), 62.3 (d)

Scheme 1 .
Scheme 1. Synthesis of the DBO carboxylic acid 4. Conditions used: a) Ag + in H 2 O/THF for 16 h, b) MTAD in CH 2 Cl 2 at 0 °C, c) hydrogenation with Pd/C in ethanol for 16 h, d) KOH in ipropanol at 80 °C for 2 days, subsequent standing under air for 3 days.

Scheme 2 .
Scheme 2. Synthesis of DBO acetic acid 5. Conditions used: a) NaCN in DMSO at 110 °C for 24 h, b) NaOH in H 2 O/EtOH at 90 °C for 16 h.
also the fluorescence intensities and lifetimes of the parent 1 remain unaffected in this pH range.32

Figure 1 .
Figure 1.pH titration curves of 4 in a) H 2 O and b) D 2 O by steady-state (open circles, left scale) and time-resolved fluorescence spectroscopy (filled circles, right scale).

Figure 2 .
Figure 2. Fluorescence quenching of the DBO derivatives 4 (1 mM, open circles) and 5 (1 mM, filled circles) in H 2 O at pH 5.5 upon addition of host 17.The solid lines correspond to the fitting curves for a 1:1 binding model.

Scheme 1 .
Scheme 1. Proposed structures (based on tentative ICD assignments) of the cyclodextrin complexes of azoalkanes 4 and 5 at pH 7.0 and 2.0.

Figure 4 .
Figure 4. Different views of the X-ray crystal structure of the DBO acid 4.